Simulating the impact of income distribution on poverty reduction.
Hyder, Syed Kalim ; Ahmed, Qazi Masood ; Jamal, Haroon 等
The paper makes an attempt to simulate the effects of growth and
income distribution on poverty reduction in the context of Pakistan. The
exercise is carried out with the help of an Integrated Social Policy
Macro Model. Macroeconomic and social sector data including poverty and
inequality are used for the period 1979 to 2013 to estimate the model
and for simulations.
The paper attempts to explore the determinants of poverty and
inequality. The findings highlight that food inflation and sectoral wage
gap in favour of manufacturing exacerbate inequality, while progressive
taxation, investment and development expenditure on social services play
a significant role in reducing inequality. The results further specify a
positive and significant relationship between per capita GDP and income
inequality. Thus the growth process reveals structural flaws in the
distribution of income in Pakistan, where rich benefit more from the
growth and hence inequality increases. Moreover, the high poverty
elasticity with respect to inequality measure also confirms the
importance of income distribution in poverty reducing effort.
The paper also evaluates the sensitivity of growth and income
distribution on the efforts on poverty reduction. The resulting
simulations establish the insufficiency of growth alone as a vehicle for
poverty reduction, and consequently highlight the importance of equity
consideration in poverty alleviation efforts.
JEL Classification: 131, D31, D63
Keywords: Poverty Simulation, Income Inequality, Growth, Pakistan
1. INTRODUCTION
The traditional notion that has influenced the development thinking
for almost half a century is that economic growth is fundamental to the
development process, and that the objective of poverty reduction can
only be achieved by allowing the benefits of growth to ultimately
trickle down to the poor. The 'primacy of growth' paradigm is
based on the premise that high growth, through high investment, would
lead to higher employment and higher wages, and thereby reducing
poverty. The 'trickle-down' paradigm assumes that the benefits
of economic growth would, in the first round, accrue to the upper income
groups, and the ensuing consumption expenditures of these households
would, in subsequent rounds, accrue incomes to relatively lower income
households.
Importance of equity consideration in poverty alleviation efforts
has been brought out of the cold and now has re-entered the mainstream
development policy agenda in many developing countries. This is the
consequence of a deep-rooted disillusionment with the development
paradigm which placed exclusive emphasis on the pursuit of growth.
During 1990s, the proliferation of quality data on income distribution
from a number of countries has allowed rigorous empirical testing of
standing debates on the relative importance of growth and redistribution
in poverty reduction. While the debate is still inconclusive, the
majority of development economists emphasised, based on empirical
cross-country data, that an unequal income distribution is a serious
impediment to effective poverty alleviation [Ravallion (1997, 2001)].
Many researchers suggested that growth is, in practice the main tool for
fighting poverty. However, they also reiterated that the imperative of
growth for combating poverty should not be misinterpreted to mean that
"growth is all that matters". Growth is a necessary condition
for poverty alleviation, no doubt, but inequality also matters and
should also be on the development agenda.
The purpose of this paper is to supplement the debate by providing
empirical evidence from Pakistan's poverty trends. The paper
simulates the impact of inequality on poverty reduction in a
macroeconomic general equilibrium framework. Time series macro and
social data are used to explore the relevance of inequality for growth
and poverty reduction. Section 2 presents a brief review of
cross-country evidence and discusses the linkages among poverty,
inequality and growth. The results of econometric specification, which
treats inequality as a determinant of poverty reduction, are furnished
in Section 3. The proximate macroeconomic and structural determinants of
inequality are discussed in the next section. Section 5 presents the
simulation results of poverty under alternative inequality scenarios.
The last section gives concluding remarks.
2. INEQUALITY, GROWTH AND POVERTY NEXUS
The conceptual validation of the inevitability of inequality as a
by-product of growth is drawn from the Kuznet hypothesis, propounded in
1955. Kuznets (1955) argued that the income distribution within a
country was likely to vary over time with its progress from a poor
agricultural society to a rich industrial society. The hypothesis
predicted an increase in inequality during early periods of growth, and
reduction in inequality as the economy reaches a higher stage of
development. Thus, the 'primacy of growth' model assumes a
trade-off between growth and equity.
Based on cross-country studies, it is maintained that distribution
policies give rise to distortions in the economy, resulting in
inefficiencies that may be substantial enough to adversely affect the
overall well being of society. For instance, research by Kaldor (1957)
and Bourguignon (1981) suggests that the marginal propensity to save of
the rich is higher than that of the poor, implying that a higher degree
of initial inequality will yield higher aggregate savings, capital
accumulation, and growth. It is also argued that inequality within a
country is stable over time and changes too slowly to make a significant
difference in poverty reduction [Deininger and Squire 1998)]. The
conclusion drawn is that growth must precede distribution, and that the
poor will pay the price of growth in terms of inequality and poverty
until such time that growth builds up a 'reservoir' of wealth
and its benefits trickle down in sufficient measure to reduce poverty.
The 'primacy of growth' paradigm has been challenged by
empirical evidence based on rigorous testing of more recent
cross-country data, and the 'trickle-down' paradigm has been
effectively discredited. Further, it is reasoned that there does not
exist an unavoidable trade-off between growth and equity [Naschold
(2002)]. Results show that high inequality is an impediment not only to
poverty reduction, but also to growth. Initial cross-country studies,
including Birdsall, et al. (1995), found that greater initial income
inequality actually reduces future growth even after controlling for
initial levels of GDP and human capital. The robustness of these
findings has been the subject of much debate; however recent analysis
using an updated and more comparable inequality data reconfirms the
negative effects of inequality on growth [Knowles (2001)]. Low
inequality can therefore benefit the poor in two ways. By increasing
overall growth and average incomes and by letting they share more in
that growth.
It is also argued that a more equitable distribution of assets and
income is likely to strengthen aggregate market demand, expand the
economic base, and foster growth. Thus, distribution is not only a final
outcome, but also a determinant of economic growth. Given that there is
no trade-off per se between growth and equality, it follows that
distribution can be pursued as an additional policy objective to enhance
the poverty reducing effect of growth. The removal or correction of the
various anti-poor institutional constraints and policy-induced biases is
likely to actually improve market efficiency, besides promoting equity.
For instance, social policy ensuring adequate provision of education and
health services to the poor can improve their productivity and
contribution to the economy. Therefore, the conclusion drawn is that
poverty reduction is not a function of high or low growth, but rather of
distribution sensitive growth.
Policies and growth patterns that improve distribution are
therefore potentially significant additional tools in the fight against
poverty. Past changes in distribution occurred without active policy
intervention, as the focus of development policy and research was on
growth, rather than distribution issues. If, in future, development
policy makes inequality an explicit target, it will greatly enhance the
poverty reducing effect of growth.
3. INEQUALITY AS A DETERMINANT OF POVERTY
International evidence shows that the poverty elasticity of growth
depends on the specific poverty measure being used [Kakwani (1993)], the
degree of inequality of the income distribution [Revallion (1997)] as
well as the specific characteristics of growth episodes, i.e., whether
growth is inequality increasing or decreasing. As such, the degree of
poverty is postulated to be a function of two factors: the average
income level of the country and the extent of income inequality.
Formally,
P = P(Y,L(p)) (1)
Where P is a poverty measure, Y is per capita income and L(p) is
the Lorenz Curve measuring the relative income distribution. The Lorenz
Curve is based on ranking of population according to income and plotting
the cumulative proportion of income against the cumulative proportion of
population enjoying that income.
Changes in poverty can be decomposed into a growth component that
relates changes in per capita income, and an inequality component that
relates poverty to changes in inequality. In general, increases in
average income (growth) will reduce poverty. Thus, growth elasticity of
poverty ([lambda]) may be hypothesised as follows:
[lambda] = [[partial derivative]P / [partial derivative]Y x y / p]]
> 0 (2)
Measuring the effect of inequality on poverty is slightly more
complex because inequality can change in infinite manners. It is hard to
say anything general about the growth-poverty relationship when the
distribution is allowed to change during growth. Although intuitively
progressive distributional change is likely to reduce poverty, this
result cannot be generalised without additional assumption regarding the
distribution. Kakwani (1993) developed a formula for the inequality
elasticity of poverty under the assumption of an equal proportionate
change in the Lorenz curve. Under this assumption it is possible to
express the inequality elasticity of poverty (o as the elasticity of
poverty with respect to the Gini coefficient (G).
[omega] = [[partial derivative]P / [[[partial derivative]G] x [G /
P]] > 0 (3)
To establish the relationship between poverty, growth and
inequality, Pakistan's time series (1979-2013) data on per capita
income, headcount (poverty incidence or population below the poverty
line) and Gini coefficient are used to estimate the following
specification. In order to capture the asymmetric impact of Gini
coefficient on poverty, the Gini is decomposed into two variables1 by
taking the threshold of no change; the Gini coefficient that observe the
increasing (positive) trend and the Gini coefficient that observe the
declining (negative) trend. These two variables are added in the
Equation (4) instead of one time series of Gini coefficients to capture
the disproportional impact of inequality on poverty.
Log [(Poverty).sub.t] = [[alpha] + [lambda] log [(GDP).sub.t] +
[[bar.[omega].sub.1][([Gini.sup.positive]).sub.t] +
[[bar.[omega].sub.2][([Gini.sup.negative]).sub.t] + gt] + [[mu].sub.t]]
(4)
As consumption and income data are collected occasionally from
Household income and expenditure Surveys, poverty and inequality series
are interpolated before estimation. Moreover, a consistent time series
of poverty is developed to avoid the inter-temporal methodological
biases.2 The estimated results of Equation (4) are furnished below.
The results from the econometric analysis clearly indicate the
importance of income distribution in determining absolute poverty level.
The poverty elasticity with respect to Gini observing increasing trend
(positive changes) and Gini witnessing declining trend (negative
changes) is estimated as 2.30 and 2.04, while the estimated poverty
elasticity with respect to income is 0.42. The higher elasticity of
poverty with respect to Gini implies that distribution is more important
as poverty predictor than income and confirms the role of inequality in
the prevalence of and/or increase in poverty.
4. EXPLAINING INEQUALITY
Given the importance of inequality as a determinant of absolute
poverty, an attempt has been made to identify important variables that
influence the Gini coefficient, particularly factors that can be
manipulated at the policy level to affect poverty.
There is widespread consensus that macroeconomic stability is a
prerequisite for pro-poor growth. In particular, it has been found
repeatedly that high inflation (particularly above a level of about 10
percent) hurts the poor and economic growth. Therefore, inflation (food
prices) may be a good proxy for fiscal stabilisation in an economy.
A negative relationship is hypothesised between development
expenditure, especially on social services (3) and income distribution.
More public expenditure on health and education certainly increases the
human capital endowment of the poor and hence affects on the
empowerment.
A major redistribution policy is to make the tax structure
pro-poor. Therefore, it is hypothesised that there is a direct link
between progressive tax structure (4) and equity. Investments,
especially in infrastructure have a major impact on making economic
growth pro-poor. Growth in investments is essential for reducing rate of
unemployment and under-employment in the economy. Public investments by
providing infrastructure play an important role in reducing poverty and
increasing the share of people at the bottom of the income distribution.
(5)
Two elements of economic structure are considered in the analysis:
first, the manufacturing to agriculture wage (6) gap and secondly, the
manufacturing to agriculture terms of trade. (7) Keeping the economic
structure of the country, it is expected that the increase in these
ratios will worsen the income distribution and will have a positive
relationship with the Gini coefficient.
Equation 5 summarises these determinants (8) of income inequality,
while estimated results of the equation are furnished in Table 2.
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)
The determinants of income inequality in the order of estimated
magnitude of impact (elasticities) are: food prices; per capita income;
manufacturing-to-agriculture terms of trade; investment/GDP ratio;
direct/indirect tax ratio; ratio of development expenditure on social
services to GDP; and ratio of manufacturing and agricultural wages.
The results indicate that average growth worsens distribution and
is unlikely to help in reducing poverty, without explicit distribution
policies. This is evident from the fact that an increase in per capita
income also raises inequality, with a one percent increase in per capita
income raising inequality by 0.172 percent. Real wheat prices emerge as
the most important determinant of inequality as measured by magnitude of
the estimated elasticity. The analysis shows that a one percent decline
in real wheat prices lowers inequality by 0.133 percent. Raising direct
tax revenues, investment, and development expenditure on social services
by one percent each is likely to reduce inequality by 0.037, 0.167 and
0.139 percent, respectively. Further, improving agricultural wages are
also likely to reduce inequality by 0.076 percent.
5. POVERTY SIMULATIONS
The Integrated Social Policy and Macroeconomic (ISPM) model9 of the
SPDC is employed to simulate poverty and inequality under various
assumptions and scenarios. The ISPM model incorporates the social,
fiscal and macroeconomic dimensions of the economy under one
interrelated system. It provides the basic framework for analysing the
implications of numerous economic measures on the long-term development
of Pakistan's social sectors. The Poverty Module has recently
modified and Income distribution is introduced in the block after having
powerful evidence of the fact that the nature of growth in Pakistan is
'inequality-increasing' and the income distribution is an
important determinant of absolute poverty. The Poverty and Income
Distribution Block of the model consists of Equations (4) and (5) with
the specification and estimated magnitudes described above.
Table 3 presents the simulation results of various combinations of
growth and inequality to achieve the desired level of poverty. The
simulations results show that if the GDP growth rate continued to be
maintained at 6 percent per annum and measures were adopted to hold the
Gini coefficient constant at the 2012 level of 0.400, poverty incidence
would probably decline to 38.3 percent by 2020. However, with the Gini
coefficient held constant at 0.400, lower GDP growth rates of 5 and 4
percent are likely to result in a higher incidence of poverty in 2020;
39.3 and 40.2 percent respectively. Similarly, if the GDP growth rate
were assumed to be 6 percent, reducing poverty incidence to 35 percent
in 2020 would require that the Gini coefficient to be lowered to 0.35
from 0.4.
The simulation results presented in the table clearly establish the
insufficiency of growth alone as a vehicle for poverty reduction, and
consequently, the inevitability of engaging with the task of reducing
inequality.
6. CONCLUDING REMARKS
Poverty reduction has always been a priority of development policy,
albeit sometime only at the rhetorical level. The end of the 2000s
brought increased emphasis on bringing the benefits of growth to the
poor. However, growth alone is a rather blunt instrument for poverty
reduction, since the consensus of empirical work suggests that it is
distribution neutral. Along with emphasis on poverty reduction, a shift
occurred in the policy literature towards a moire favorable view of
policies to redistribute income and assets. An integration of
distributional concerns and a priority on poverty reduction could be the
basis for a new policy agenda to foster growth with equity.
This paper highlights the importance of distribution policies in
poverty reduction using Pakistan time series macroeconomic and social
data during the period 1979 to 2013. Simulation exercise is carried out
by employing SPDC integrated macroeconomic model.
Following are the main findings of this research. First, the
poverty elasticity with respect to Gini coefficient is statistically
significant and also the magnitude is relatively high as compared with
poverty elasticity of growth. Second, the study found inflation,
sectoral wage gap, and terms of trade in favor of manufacturing as the
significant positive correlates of inequality, while progressive
taxation, investment and development expenditure on social services are
negatively impacting on inequality. Third, the simulation exercise in a
general equilibrium framework clearly demonstrates that a high GDP
growth rate, without accompanying equity-promoting policy shifts, is by
itself unlikely to reduce the incidence of poverty.
Finally, it is true that redistribution often has limited potential
and that growth is a necessary condition for poverty reduction. Yet the
level of inequality and change therein, still matters. This is because
the level of inequality affects the degree of poverty as well as growth
elasticity of poverty. Further, low level of inequality contributes for
an acceleration of poverty reduction for a given level of growth. For
these reasons, inequality still mattes, and the search for effective
policies for reducing inequality, or at least prevent them from rising,
should be an integral part of the development agenda.
APPENDIX--A
INTEGRATED SOCIAL POLICY AND MACROECONOMIC MODEL
Social Policy and Development Centre (SPDC) has developed one of
the pioneer models which can be used as an effective planning tool for
social sector development. The Integrated Social Policy and
Macroeconomic (ISPM) model integrates the social, fiscal and
macroeconomic dimensions of the economy under one interrelated system.
It provides the basic framework for analysing the implications of
numerous economic measures on the long-term development of
Pakistan's social sectors. Recently the ISPM model incorporated the
changes in Pakistan's economy by endogenising both interest rate
and exchange rate variables.
The model is highly disaggregated and covers all three levels of
government. It is capable of predicting outcomes in considerable detail,
even at the level of individual social service provision. The ability to
disaggregate the model at the provincial level in terms of revenues and
expenditures on social services (e.g., schools, hospitals, doctors,
teachers, enrolments, etc.) is helpful in analysing the impact of
related initiatives on the macro economy and social development.
The ISPM model is based on consistent national level data from 1973
onwards and is estimated by single equation regression techniques. It
consists of 409 equations, of which 172 are behavioral and the rest are
identities. These equations are subsumed into 18 interrelated blocks.
The blocks, along with their size in terms of equations and identities,
are listed in Table A. 1.
Although the model is broadly Keynesian in spirit, the
specification of individual blocks and equations is based on a pragmatic
approach and also captures the non-market clearing aspects of
Pakistan's economy. Thus, the macroeconomic block is essentially
supply driven. In addition, the social sector indicators are also
resource determined.
[ILLUSTRATION OMITTED]
The model has dynamic specifications which vary across the blocks.
In some cases, the linkage is simultaneous and in some cases it is
recursive. Examples include the linkages between the macro-production
and input blocks; the production and expenditure blocks; the fiscal
revenues and expenditure blocks; and the macro production, poverty and
inequality blocks. The broad links (see Chart A.l) of the model can be
traced as follows.
Macro--Public Finance
The key link here traces the impact of developments in the
macroeconomy on the growth of the tax bases (including divisible pool
taxes) and thus affects the fiscal status of different governments.
Public Finance--Social Sector Development
The availability of resources, both external and internal,
determines the level of development and recurring outlays to social
sectors by different levels of government, particularly provincial and
local.
Social Sector Development--Macroeconomy
Higher output of educated workers and their entry into the labour
force raises the human capital stock and could contribute to
improvements in productivity and a higher growth rate of output in the
economy. Similarly, an improvement in public health standards may also
have a favorable impact on production.
Public Finance--Macroeconomy
The level of government expenditure could exert a demand side
effect on national income, while the size of the overall budget deficit
of the federal and provincial governments influences the rate of
monetary expansion and consequently the rate of inflation in the
economy.
Social Sector Development--Public Finance
A vital link in the model is between the rate of social sector
development and the state of public finances. Higher social sector
development implies higher recurring expenditures of provincial
governments, which are consequently reflected in the budget deficit,
level of debt stock and debt servicing of provincial governments.
[ILLUSTRATION OMITTED]
Macro Economy--Social Sector Development
Macro and other socio-economic changes affect the demand for social
sector facilities such as schools and hospitals, and thus influence the
level of social sector outputs.
Apart from these broad linkages among different modules, there are
also links between different blocks within each module (see Chart A.2).
An example of a major linkage within the macro module is the
two-way linkage to and from the macro-production block and macro-input
blocks. This link is due to the dependence of sectoral value added to
the factors of production and input demand functions on the value of
production. Macro production determines macro expenditure, just as
private consumption is influenced by income.
The two-way link between the macro-production block and the trade
block is due to the fact that the value of imports and exports
determines and is determined by economic production activity. The trade
gap affects the level of money supply.
Important linkages in the fiscal module consist of the simultaneous
dependence of revenues and expenditures of various levels of government.
Non-tax receipts of governments have been made a function of the
recurring expenditure on particular services via cost recovery ratios.
Similarly, the level of government expenditure is affected by the
government's level of resource generation.
Important vertical links between levels of government include
fiscal transfers in the form of divisible pool transfers and
non-development grants (in line with the feasible level of
decentralisation) from provincial to local governments. The link between
the budget deficits of the federal and provincial governments and their
revenues and expenditures is obvious.
Forecasting and Policy Analysis Tool
Given the richness of its structure and the complex web of
interrelationships and interactions it embodies, the 1SPM model can be
used both as a forecasting tool for the medium and long term, and for
undertaking policy simulations to analyse the consequences of particular
policy actions by the government.
For example, if the federal government decides to pursue a policy
of higher tax mobilisation and opts for a rigorous fiscal effort, the
model can forecast the impact, not only on federal finances, but also on
the fiscal status of the provincial governments. In this scenario, it
could also forecast key macroeconomic magnitudes such as growth in the
gross domestic product, social development, budget deficit, changes in
income inequality and the inflation rate.
The model can also perform simulations to find the relative
strength of different policy options for a specific objective. In the
case of the macro economy, it can provide the impact of different policy
options on:
* short and medium-term projections of the growth of important
sectors (agriculture, manufacturing, construction, electricity and gas
distribution);
* short and medium-term projections of the growth of GDP, GNP, per
capita income;
* factor input (e.g., capital and labor) demand; and
* short and medium-term projections of the public and private
investment in various sectors of the economy.
In the case of pubic finance, it can:
* provide short and medium-term projections of the quantum of
revenue transfers to the provincial governments by the federal
government under different scenarios;
* determine the impact of different rates and patterns of economic
growth on provincial tax bases and revenues; and
* determine the impact of changes in provincial expenditure
priorities on fiscal status, levels of service provision and the overall
macro economy.
In the case of social development, it can determine the impact on:
* poverty reduction strategy related expenditures;
* social sector expenditures by provincial governments on income
inequality that further changes the poverty rate;
* education expenditures by provincial governments on sectoral
inputs (schools, teachers), enrolments, outputs, entry into the labour
force and literacy rates;
* health expenditures by provincial governments on sectoral inputs
(beds, rural health centres, doctors, nurses, paramedics) and on the
health status of the population; and
* higher levels of resource mobilisation by provincial governments
on federal transfers, sectoral levels of expenditure and fiscal status.
Income Inequality and Poverty Block
An important aspect of the SPDC's macro model is the
incorporation of the poverty and inequality block. In this block, the
linkage of macro, public finance and human development variables with
the measure of income inequality (Gini Coefficient) is developed, which
also helps in determining poverty. This is one of the pioneer works in
the economic literature of developing countries that explores the impact
of economic growth and government expenditures on income inequality and
poverty. The complete linkages between growth, income distribution and
poverty are shown in Chart A3.
[ILLUSTRATION OMITTED]
Table A. 1
Integrated Social Policy and Macro Model (ISPM)
Total Equations Identities
Block A Production Block 27 11 16
Block B Input Block 37 16 21
Block C Aggregate Demand Block 34 20 14
Block D Trade and Balance of 19 11 8
Payments Block
Block E Monetary and Prices Block 10 7 3
Block F Federal Revenue Block 12 4 8
Block G Federal Expenditure Block 16 9 7
Block H Sub-National Revenue Block 26 11 15
Block I Sub-National Expenditure 32 22 10
Block
Block J Debt and Budget Deficit 12 2 10
Block
Block K Education Block 47 24 23
Block L Human Capital Index Block 16 5 11
Block M Health Block 27 18 9
Block N Public Health Index Block 4 3 1
Block O Human Development Index 7 0 7
Block
Block P Poverty and Income 12 3 9
Inequality Block
Block Q Goals Block 63 6 57
Block R Costing and Financing Block 8 0 8
Total 409 172 237
REFERENCES
Birdsall, Nancy, David Ross, and Richard R. Sabot (1995) Inequality
and Growth Reconsidered: Lessons from East Asia. World Bank Economic
Review 9:3, 477-508.
Bourguignon, F. (1981) Pareto Superiority of Unegalitarian
Equilibrium in Stiglitz's Model of Wealth Distribution with Convex
Saving Function. Econometrica 49, 1469-75.
Deininger Klaus, and Lyn Squire (1998) New Ways of Looking at Old
Issues: Inequality and Growth. Journal of Development Economics 57:2,
259-287.
Jamal, Haroon (2006) Does Inequality Matter for Poverty Reduction?
Evidence from Pakistan's Poverty Trends. The Pakistan Development
Review 45:3.
Kakwani, Nanak (1993) Poverty and Economic Growth with an
Application to Cote d'Ivoire. Review of Income and Wealth 39:2.
Kaldor, N. (1957) A Model of Economic Growth. Economic Journal 67.
Knowles, Stephen (2001) Inequality and Economic Growth: The
Empirical Relationship reconsidered in the Light of Comparable Data.
Paper for UNU/Wider Development Conference on Growth and Poverty, WIDER,
Helsinki.
Kuznets, Simon (1955) Economic Growth and Income Inequality.
American Economic Review 45, 1-28.
Naschold, Felix (2002) Why Inequality Matters for Poverty? UK
Department for International Development (DFID), UK. (Briefing Paper
Number 2).
Ravallion, Martin (2001) Growth, Inequality and Poverty: Looking
Beyond Averages. World Development 29:11, 1803-1815.
Ravallion, Martin (1997) Can High Inequality Developing Countries
Escape Absolute Poverty? World Bank, Washington, D.C. (World Bank
Working Paper Number 1775).
Syed Kalim Hyder <skhbl@le.ac.uk> is affiliated with the
Department of Economics, University of Leicester, UK. Qazi Masood Ahmed
<qmasood@iba.edu.pk> is Professor and Director, Centre for
Business and Economics Research, Institute of Business Administration,
Karachi Haroon Jamal <haroonjamal@ hotmail.com> is Visiting
Research Fellow, Institute of Business Administration, Karachi and
Technical Advisor, Social Policy Development Centre, Karachi.
Authors' Note: First version of the paper was completed when
all three authors were associated with the Social Policy and Development
Centre. Authors are grateful to Dr Hafiz Pasha for his guidance and
support for this paper
(1) Ideally Atkinson class of measures or extended Gini should be
used with high value of inequality aversion parameters to represent the
level of society concern about inequality. Nonetheless, this was not
possible due to non-availability of time-series raw data.
(2) The data and methodological details for interpolation and
construction of consistent poverty estimates are provided in Jamal
(2006).
(3) This is included as percent of GDP.
(4) The ratio of Direct taxes to Indirect taxes is used as a proxy
for progressivity in tax structure.
(5) Some other possible candidates for explaining inequality, like
economic and food subsidies, remittances, unemployment rate etc. were
also tested, but not turned out statistically significant.
(6) Sectoral wage is computed as the sectoral value added divided
by sectoral labour force.
(7) This is the ratio of manufacturing implicit GDP deflator to
that of agriculture implicit GDP deflator.
(8) Data on the per capita income, investment, term of trade
between agriculture and manufacturing and food prices are taken from
various issues of Pakistan Economic Survey. Relative wages are taken
from various issues of Labour Force Survey. Development expenditures,
direct tax and indirect taxes are collected from various issues of
Federal Budget in Brief.
(9) The detail description of the model with various linkages is
provided in the Appendix A.
Table 1
Determinants of Poverty
Dependent Variable--Log (Poverty Incidence--Headcount)
Explanatory
Variables Coefficient t-Statistic Significance
GDP Per Capita -0.42 -2.26 0.03
GINI (High Changes) 2.30 4.13 0.00
GINI (Low Changes) 2.04 3.69 0.00
Time Trend 0.01 2.46 0.02
Constant 6.60 3.36 0.00
R-squared 0.95 F-Statistic 103
Adjusted R-squared 0.94 Probability 0.000
(F-Statistics)
Durbin-Watson stat 1.60 Number of 40
Observations
[Q.sub.(1)] 0.81 Jarque-Bera 2.01
(0.80) (0.36)
[Q.sup.2.sub.(1)] 0.50 [LM.sub.(1)] 3.00
(0.48) (0.12)
[ARCH.sub.(1)] 0.44
(0.51)
Notes: All variables are in logarithmic form and statistically
significant.
Equation also contains a dummy variable for the year 2011 and 2012 due
to large residual effect.
LM and ARCH tests are applied and found no evidence of serial
correlation.
Wald test is applied to test the hypothesis that Gini has symmetric
impact on poverty. The hypothesis is rejected by F-test (F-value 22.75
with probability of 0.00).
Table 2
Determinants of Inequality
Dependent Variable: Log (Gini Coefficient)
Explanatory
Variables Coefficient t-Statistic Significance
Per Capita GDP 0.172 3.81 0.00
Real Price of Wheat 0.133 4.82 0.00
Wage Gap 0.076 2.72 0.01
Direct to Indirect
Tax Ratio -0.037 -1.87 0.07
Development
Expenditure on
Social Services -0.139 -8.07 0.00
Investment -0.167 -4.59 0.00
Constant (Intercept) -1.805 -3.50 0.00
R-squared 0.922 F-statistic 39.321
Adjusted R-squared 0.898 Probability (F-St.) 0.00
Durbin-Watson stat 1.409 Number of Observations 40
[Q.sub.(1)] 0.227 Jarque-Bera 1.299
(0.14) (0.52)
[Q.sup.2.sub.(1)] 0.104 [LM.sub.(1)] 2.127
(0.49) (0.16)
[ARCH.sub.(1)] 0.452
(0.51)
Notes: Variables (except dummy) are used after Logarithmic
transformation.
LM and ARCH tests are applied and found no evidence of serial
correlation.
Three dummy variables are also used in the equation to captures the
extreme point estimates.
Table 3
Simulation of Poverty Incidence with Alternative Growth and Inequality
Scenario
Gini Coefficient Scenario
0.400 0.385 0.350
GDP Growth Rate Scenario
6% 5% 4% 6% 5% 4% 6% 5% 4%
Year
2013 = 40.5 40.5 40.5 40.5 40.5 40.5 40.5 40.5 40.5
Base
2015 39.3 39.5 39.6 38.1 38.3 38.4 35.6 35.7 35.8
2017 39.0 39.5 39.8 37.8 38.3 38.6 35.3 35.6 35.9
2020 38.3 39.3 40.2 37.2 38.1 39.0 34.7 35.4 36.3
Source: SPDC Macroeconomic Model Simulations.
